Cutting rolls for a player piano

I’m looking to try and make my own rolls for a player piano I acquired. Given the size of the rolls, it may be possible to do so on one of the machines like the vinyl cutter.

Ideally, I’d like to write a processor to convert the MIDI files straight to G Code. If that fails, I could output a really long image to do the cut from. Can the machines take raw G Code, and if so does anyone know anything about that process?

The other potentially tricky part is the length of the roll. These machines are nice in that they can work in sections of a continuous roll, but the “long play” rolls can still go upwards of 100 feet in length.

Any thoughts?

-Tails

I would suggest starting with a service first like,
http://www.sierramusicrolls.com/piano_midi.html

That way you can learn about the common issues in just converting the midi. Then maybe tackle making your own with the knowledge learned. Also, make sure you understand the tolerances for the roll. Because vinyl cutters do not roll truly strait. They will walk one direction while cutting. You can see this easily by just feeding vinyl about 5 or 6 feet through it. I would imagine even a well aligned piece to walk inches across 100 feet of length. You can account for some of that by cutting on larger media and connection sections together rather than one run.

2 Likes

This would be a fun project. How I would do it:

Build a reel-to-reel drive with a stepper. Use a low-wattage burning laser (these are inexpensive) on a single axis drive. Advance the web in steps and burn the holes.

1 Like

I’d actually reached out to a couple already. For the first several I plan on using one of them to get the hang of the process anyway (so far haven’t heard back from any, but follow up with more today and that particular service isn’t one I had so I’ll give them a whirl too)

This is really good to know and a big problem. Might be better to make my own CNC for it them so I can machine it to some tolerance.

This is something to consider knowing about the drift on the vinyl cutter.

The trick with doing a reel-to-reel is that your take-up reel will be getting larger in diameter as you go, and you’ll have to decrease your rotation angle to compensate.

But it’s otherwise like a line printer.

1 Like

an easy way to deal with that are leveling rollers between the laser head and the spool out and take ups. Put the encoders on those rollers and it will always know how much is actually feeding in/out regardless of spool size. Will help with keeping things taught as well

3 Likes

I have need to make a reel-to-reel RFID tag encoder, which is half of this battle. If you have any interest in collaborating on this, let’s talk.

In the corner of the same room where the vinyl cutter is there are several rolls of wide plotter paper. These were given to DMS along with the old plotters that are no longer used. Now these rolls of paper are about 4 feet wide and over 100 feet long so you woud have to come up with some sort of paper slitter to cut it to size. So, there is your source for paper.

2 Likes

If all you wanted is some “software,” rolls can be found on eBay; prices vary tremendously.

@engpin @talkers

Did you guys ever figure out something which worked?

@bill not sure what Jayson did in the end.

One idea - the Silhouette Cameo can cut 12" rolls up to 10’ long with rollers, not needing a mat/backing sheet.

Using a small laser and custom rollers is a great idea @mdredmond & @hon1nbo

3 Likes

A player piano roll is 11.25" wide and typically 25 to 90 feet long.

http://www.pianorolls.co.uk/Anatomy%20of%20a%20Piano%20Roll.htm#:~:text=A%20piano%20roll%20consists%20of,feet%20for%20the%20longest%20rolls.

1 Like

So the Cameo can handle the width, and even trim 12" rolls. then tape splice the rolls together if you have a longer song.

This sounds like a good chance to use your ballistic art skills to perforate the roll? Music in the Key of all notes! :rofl:

1 Like

I ran across this thread while searching the Internet to find if a commercial piano roll perforating machine is still available. I found your discussion interesting, and I also found someone on YouTube showing video of a perforating machine they had developed. I wondered if anyone who posted in this thread has seen that. It looks pretty much like what you folks were envisioning. According to YouTube, this was posted 10 years ago. Here’s the YouTube link, in case some of you may have missed it. https://www.youtube.com/watch?v=X8ManoX9mUs He uses MIDI files to control the perforator. I like how he set this up, except for the fact he is utilizing a circular punch, whereas tracker bar holes are rectangular. Also, his punching method leaves something to be desired because multiple overlapping punches for extended note length often leave “hanging chad” type punch failures that would have to be manually excised. Other than that, though, it’s a pretty well thought out perforator for making piano rolls.

4 Likes

Very cool,
Thanks for sharing.

I wonder if a low-wattage laser would do this effectively?

After the original discussion had faded away, I found this video about how QRS made rolls:

I do not have room for an actual player piano, but I would love to have some kind of a gizmo which “plays” a roll to control a MIDI keyboard…

1 Like

I found out with my Trotec Laser, there is a way to use the rotary attachment for semi continuous laser cutting. You effectively make the rotary a pinch roller and it pulls the paper as it cuts. The hack has to be run as an engrave so as to only roll one direction.

The X-Tool Diode laser now has a feed option on their new model as well.

2 Likes

Hi, Nick. As you’re probably already aware, all Digital Audio Workstation (DAW) software programs feature a “piano roll” view which is pretty nifty and allows for more control than entering notes on staffs, and of course allows for playback on a MIDI keyboard. A great way to experiment with this is by downloading and installing Cakewalk by Band Lab, which is free software which used to be known as Sonar Platinum and sold for about $400. I suspect, though, that what you are suggesting is using a tracker bar mechanism and actual player piano rolls to play through a MIDI keyboard. That would certainly be feasible, but the cost could be somewhat high. I used to rebuild player pianos for a living, and found that one of the best places to find player piano mechanisms on the cheap was at old churches. That’s because when a player piano stopped working, most owners either sold the piano cheaply, or gave it to a church, because the cost to rebuild the mechanism was prohibitive. I was up inside a bell tower of a church and actually found two player piano mechanisms that had been stored there. In looking around the church, I also found the two pianos they had been removed from, so I proposed a deal wherein I would take the pianos and the player mechanisms in exchange for tuning and repairing another piano which they had but were unable to use because some demented prankster had snapped off all 88 of the piano hammers. Another good place to look for a player piano mechanism is Facebook Marketplace, where you can probably find an old player piano that’s cheap, or even free for the taking. Chances are pretty good that you would need to restore the mechanism so that it is functional, but that isn’t difficult and the materials needed are fairly inexpensive. Aside from the player mechanism, you will also need a vacuum source. Older player pianos had foot pumped pedals which operated bellows to provide the vacuum, but you could use an electric vacuum pump made for player piano conversions, or even a vacuum cleaner for that matter. Once you have the mechanism functioning, you would need to convert the mechanical actions, which would ordinarily activate piano keys, into digital signals that would feed to the MIDI keyboard via a standard or USB MIDI cable. Another option, when using a player piano mechanism, would be to entirely remove or bypass the key activating mechanisms, which are basically operated by pneumatic pouches connected to the tracker bar by rubber tubing. You could, for example, insert fiberoptic strands through each of the tubes, placing one end very close to the tracker bar, and the opposite end could be used to activate a photocell or photoresistor to complete a circuit. An LED light bar shining on the roll paper as it moves across the tracker bar would activate each fiberoptic strand when its corresponding slot opens in the paper roll. This would of course eliminate the need of rebuilding the entire mechanism, although replacing the rubber tubing would be advised.

1 Like

The diameter of the take-up spool effectively increases by two paper thicknesses per turn. If the reference is the position of the take-up spool while a roll is being cut, the problem essentially solves itself. Trying to compensate for this effect with any other feed mechanism would be a nightmare.